diff options
| -rw-r--r-- | src/fortrancode.l | 50 | ||||
| -rw-r--r-- | src/fortranscanner.l | 295 |
2 files changed, 295 insertions, 50 deletions
diff --git a/src/fortrancode.l b/src/fortrancode.l index fb2b641..796448a 100644 --- a/src/fortrancode.l +++ b/src/fortrancode.l @@ -94,6 +94,7 @@ static int g_inputPosition; //!< read offset during parsing static int g_inputLines; //!< number of line in the code fragment static int g_yyLineNr; //!< current line number static bool g_needsTermination; +static bool g_isFixedForm; static bool g_insideBody; //!< inside subprog/program body? => create links static const char * g_currentFontClass; @@ -109,6 +110,44 @@ static bool g_includeCodeFragment; static char stringStartSymbol; // single or double quote +// simplified way to know if this is fixed form +// duplicate in fortranscanner.l +static bool recognizeFixedForm(const char* contents) +{ + int column=0; + bool skipLine=FALSE; + + for(int i=0;;i++) { + column++; + + switch(contents[i]) { + case '\n': + column=0; + skipLine=FALSE; + break; + case ' ': + break; + case '\000': + return FALSE; + case 'C': + case 'c': + case '*': + if(column==1) return TRUE; + if(skipLine) break; + return FALSE; + case '!': + if(column>1 && column<7) return FALSE; + skipLine=TRUE; + break; + default: + if(skipLine) break; + if(column==7) return TRUE; + return FALSE; + } + } + return FALSE; +} + static void endFontClass() { if (g_currentFontClass) @@ -837,6 +876,15 @@ IGNORE (IMPLICIT{BS}NONE|CONTAINS|WRITE|READ|ALLOCATE|DEALLOCATE|SIZE) codifyLines(yytext); endFontClass(); } + +<*>^[Cc*].* { // normal comment + if(! g_isFixedForm) REJECT; + + startFontClass("comment"); + codifyLines(yytext); + endFontClass(); + } + /*------ preprocessor --------------------------------------------*/ <Start>"#".*\n { startFontClass("preprocessor"); codifyLines(yytext); @@ -884,6 +932,7 @@ IGNORE (IMPLICIT{BS}NONE|CONTAINS|WRITE|READ|ALLOCATE|DEALLOCATE|SIZE) /*===================================================================*/ + void resetFortranCodeParserState() {} void parseFortranCode(CodeOutputInterface &od,const char *className,const QCString &s, @@ -901,6 +950,7 @@ void parseFortranCode(CodeOutputInterface &od,const char *className,const QCStri g_code = &od; g_inputString = s; g_inputPosition = 0; + g_isFixedForm = recognizeFixedForm((const char*)s); g_currentFontClass = 0; g_needsTermination = FALSE; if (endLine!=-1) diff --git a/src/fortranscanner.l b/src/fortranscanner.l index 5722236..89d5473 100644 --- a/src/fortranscanner.l +++ b/src/fortranscanner.l @@ -112,10 +112,18 @@ static const char *directionStrs[] = static ParserInterface *g_thisParser; static const char * inputString; static int inputPosition; +static bool isFixedForm; static QCString inputStringPrepass; ///< Input string for prepass of line cont. '&' static unsigned int inputPositionPrepass; static int lineCountPrepass = 0; +struct CommentInPrepass { + int column; + QCString str; + CommentInPrepass(int column, QCString str) : column(column), str(str) {} +}; +static QList<CommentInPrepass> comments; + #define MAX_INCLUDE_DEPTH 10 YY_BUFFER_STATE include_stack[MAX_INCLUDE_DEPTH]; int include_stack_ptr = 0; @@ -174,9 +182,12 @@ static bool endScope(Entry *scope, bool isGlobalRoot=FALSE); static QCString getFullName(Entry *e); static bool isTypeName(QCString name); static void resolveModuleProcedures(QList<Entry> &moduleProcedures, Entry *current_root); -static int getAmpersandAtTheEnd(const char *buf, int length); +static int getAmpersandAtTheStart(const char *buf, int length); +static int getAmpOrExclAtTheEnd(const char *buf, int length); +static void truncatePrepass(int index); static void pushBuffer(QCString &buffer); static void popBuffer(); +static void extractPrefix(QCString& text); //----------------------------------------------------------------------------- #undef YY_INPUT @@ -257,37 +268,45 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA /*-----------------------------------------------------------------------------------*/ -<*>^.*"&"{BS}(!.*)?\n { // ampersand is somewhere in line +<*>^.*\n { // prepass: look for line continuations + //fprintf(stderr, "---%s", yytext); - int index = getAmpersandAtTheEnd(yytext, yyleng); - if(index<0){ // ampersand is not line continuation - if(YY_START == Prepass) { // last line in "continuation" - inputStringPrepass+=(const char*)yytext; - pushBuffer(inputStringPrepass); - yy_pop_state(); - } else { // simple line - REJECT; - } - } else { // line with continuation - inputStringPrepass+=(const char*)yytext; - lineCountPrepass ++; - - // replace & with space and remove the following chars - int length = inputStringPrepass.length(); - inputStringPrepass[length-yyleng+index] = ' '; - inputStringPrepass.truncate(length-yyleng+index+1); - if(YY_START != Prepass) - yy_push_state(Prepass); - } - } + int indexStart = getAmpersandAtTheStart(yytext, yyleng); + int indexEnd = getAmpOrExclAtTheEnd(yytext, yyleng); + if (indexEnd>=0 && yytext[indexEnd]!='&') //we are only interested in amp + indexEnd=-1; + + if(indexEnd<0){ // ----- no ampersand as line continuation + if(YY_START == Prepass) { // last line in "continuation" + inputStringPrepass+=(const char*)yytext; + if(indexStart>=0) inputStringPrepass[yyleng-indexStart]=' '; + // @todo: remove all symbols instead of replacing W blank? + + pushBuffer(inputStringPrepass); + yy_pop_state(); + } else { // simple line + REJECT; + } + + } else { // ----- line with continuation + if(YY_START != Prepass) { + comments.setAutoDelete(TRUE); + comments.clear(); + yy_push_state(Prepass); + } + + inputStringPrepass+=(const char*)yytext; + lineCountPrepass ++; + + // replace & with space and remove following comment if present + int length = inputStringPrepass.length(); + truncatePrepass(length-yyleng+indexEnd); + } -<Prepass>^.*\n { - inputStringPrepass+=(const char*)yytext; - pushBuffer(inputStringPrepass); - yy_pop_state(); } + /*------ ignore strings */ <*>"\\\\" { /* ignore \\ */} <*>"\\\""|\\\' { /* ignore \" and \' */} @@ -472,7 +491,7 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA BEGIN(TypedefBody); } <TypedefBody>^{BS}"end"{BS}"type"({BS_}{ID})?{BS}/(\n|!) { /* end type definition */ - //cout << "=========> got typedef end "<< endl; + //printf("=========> got typedef end \n"); if (!endScope(current_root)) yyterminate(); yy_pop_state(); @@ -639,23 +658,7 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA // TYPE_SPEC is for old function style function result result= yytext; result= result.stripWhiteSpace(); - int prefixIndex = 0; - int curIndex = 0; - bool cont = TRUE; - const char* pre[] = {"RECURSIVE","PURE","ELEMENTAL"}; - while(cont) - { - cont = FALSE; - for(unsigned int i=0; i<3; i++) - { - if((prefixIndex=result.find(pre[i], curIndex, FALSE))==0) - { - result.remove(0,strlen(pre[i])); - result.stripWhiteSpace(); - cont = TRUE; - } - } - } + extractPrefix(result); //fprintf(stderr, "===%s\n", (const char*)result); current->type = result; yy_push_state(SubprogPrefix); @@ -667,9 +670,12 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA BEGIN(Subprog); } -<Start,ModuleBody,SubprogBody,InterfaceBody,ModuleBodyContains,SubprogBodyContains>^{BS}{SUBPROG}{BS_} { +<Start,ModuleBody,SubprogBody,InterfaceBody,ModuleBodyContains,SubprogBodyContains>^{BS}({PREFIX}{BS_})?{SUBPROG}{BS_} { // Fortran subroutine or function found - addSubprogram(yytext); + result= yytext; + result= result.stripWhiteSpace(); + extractPrefix(result); + addSubprogram(result); yy_push_state(Subprog); } @@ -788,6 +794,7 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA } <*>. { //debugStr+=yytext; + //printf("I:%c\n", *yytext); } // ignore remaining text /**********************************************************************************/ @@ -796,12 +803,50 @@ PREFIX (RECURSIVE{BS_}|PURE{BS_}|ELEMENTAL{BS_}){0,2}(RECURSIVE|PURE|ELEMENTA %% //---------------------------------------------------------------------------- -static int getAmpersandAtTheEnd(const char *buf, int length) +static void extractPrefix(QCString &text) { + int prefixIndex = 0; + int curIndex = 0; + bool cont = TRUE; + const char* pre[] = {"RECURSIVE","PURE","ELEMENTAL"}; + while(cont) + { + cont = FALSE; + for(unsigned int i=0; i<3; i++) + { + if((prefixIndex=text.find(pre[i], curIndex, FALSE))==0) + { + text.remove(0,strlen(pre[i])); + text.stripWhiteSpace(); + cont = TRUE; + } + } + } +} + +static int getAmpersandAtTheStart(const char *buf, int length) +{ + for(int i=0; i<length; i++) { + switch(buf[i]) { + case ' ': + case '\t': + break; + case '&': + return i; + default: + return -1; + } + } + return -1; +} + +/* Returns ampersand index, comment start index or -1 if neither exist.*/ +static int getAmpOrExclAtTheEnd(const char *buf, int length) { // Avoid ampersands in string and comments int parseState = Start; char quoteSymbol = 0; int ampIndex = -1; + int commentIndex = -1; for(int i=0; i<length && parseState!=Comment; i++) { @@ -824,6 +869,7 @@ static int getAmpersandAtTheEnd(const char *buf, int length) break; case '!': parseState = Comment; + commentIndex = i; break; case ' ': // ignore whitespace case '\t': @@ -837,7 +883,138 @@ static int getAmpersandAtTheEnd(const char *buf, int length) } } } - return ampIndex; + + if (ampIndex>=0) + return ampIndex; + else + return commentIndex; +} + +/* Although comments at the end of continuation line are grabbed by this function, +* we still do not know how to use them later in parsing. +*/ +void truncatePrepass(int index) +{ + int length = inputStringPrepass.length(); + for (int i=index+1; i<length; i++) { + if (inputStringPrepass[i]=='!') { // save comment + //printf("-----SAVE----- %d:%s", i, (const char*)inputStringPrepass.right(length-i)); + struct CommentInPrepass *c=new CommentInPrepass(index, inputStringPrepass.right(length-i)); + comments.append(c); + } + } + inputStringPrepass[index] = ' '; + inputStringPrepass.truncate(index+1); +} + +// simplified way to know if this is fixed form +// duplicate in fortrancode.l +static bool recognizeFixedForm(const char* contents) +{ + int column=0; + bool skipLine=FALSE; + + for(int i=0;;i++) { + column++; + + switch(contents[i]) { + case '\n': + column=0; + skipLine=FALSE; + break; + case ' ': + break; + case '\000': + return FALSE; + case 'C': + case 'c': + case '*': + if(column==1) return TRUE; + if(skipLine) break; + return FALSE; + case '!': + if(column>1 && column<7) return FALSE; + skipLine=TRUE; + break; + default: + if(skipLine) break; + if(column==7) return TRUE; + return FALSE; + } + } + return FALSE; +} + +/* This function assumes that contents has at least size=length+1 */ +static void insertCharacter(char *contents, int length, int pos, char c) +{ + // shift tail by one character + for(int i=length; i>pos; i--) + contents[i]=contents[i-1]; + // set the character + contents[pos] = c; +} + +/* change comments and bring line continuation character to previous line */ +static const char* prepassFixedForm(const char* contents) +{ + int column=0; + int prevLineLength=0; + int prevLineAmpOrExclIndex=-1; + bool emptyLabel=TRUE; + int newContentsSize = strlen(contents)+2; // \000 and one spare character (to avoid reallocation) + char* newContents = (char*)malloc(newContentsSize); + + for(int i=0, j=0;;i++,j++) { + if(j>=newContentsSize-1) { // check for one spare character, which may be eventually used below (by &) + newContents = (char*)realloc(newContents, newContentsSize+1000); + newContentsSize = newContentsSize+1000; + } + + column++; + char c = contents[i]; + switch(c) { + case '\n': + prevLineLength=column; + prevLineAmpOrExclIndex=getAmpOrExclAtTheEnd(&contents[i-prevLineLength+1], prevLineLength); + column=0; + emptyLabel=TRUE; + newContents[j]=c; + break; + case ' ': + newContents[j]=c; + break; + case '\000': + newContents[j]='\000'; + return newContents; + case 'C': + case 'c': + case '*': + emptyLabel=FALSE; + if(column==1) + newContents[j]='!'; + else + newContents[j]=c; + break; + default: + if(column==6 && emptyLabel) { // continuation + newContents[j]=' '; + + if(prevLineAmpOrExclIndex==-1) { // add & just before end of previous line + insertCharacter(newContents, j+1, (j+1)-6-1, '&'); + j++; + } else { // add & just before end of previous line comment + insertCharacter(newContents, j+1, (j+1)-6-prevLineLength+prevLineAmpOrExclIndex, '&'); + j++; + } + } else { + newContents[j]=c; + emptyLabel=FALSE; + } + break; + } + } + return newContents; } static void pushBuffer(QCString& buffer) @@ -850,11 +1027,12 @@ static void pushBuffer(QCString& buffer) include_stack[include_stack_ptr++] = YY_CURRENT_BUFFER; yy_switch_to_buffer(yy_scan_string(buffer)); - //fprintf(stderr, "--POP--%s", (const char *)buffer); + //fprintf(stderr, "--PUSH--%s", (const char *)buffer); buffer = NULL; } static void popBuffer() { + //fprintf(stderr, "--POP--"); include_stack_ptr --; yy_delete_buffer( YY_CURRENT_BUFFER ); yy_switch_to_buffer( include_stack[include_stack_ptr] ); @@ -1454,6 +1632,19 @@ static void parseMain(const char *fileName,const char *fileBuf,Entry *rt) inputFile.setName(fileName); if (inputFile.open(IO_ReadOnly)) { + isFixedForm = recognizeFixedForm(fileBuf); + + if (isFixedForm) { + printf("Prepassing fixed form of %s\n", yyFileName.data()); + //printf("---strlen=%d\n", strlen(fileBuf)); + //clock_t start=clock(); + + inputString = prepassFixedForm(fileBuf); + + //clock_t end=clock(); + //printf("CPU time used=%f\n", ((double) (end-start))/CLOCKS_PER_SEC); + } + yyLineNr= 1 ; yyFileName = fileName; msg("Parsing file %s...\n",yyFileName.data()); @@ -1484,6 +1675,10 @@ static void parseMain(const char *fileName,const char *fileBuf,Entry *rt) rt->program.resize(0); delete current; current=0; moduleProcedures.clear(); + if (isFixedForm) { + free((char*)inputString); + inputString=NULL; + } inputFile.close(); } |